Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.OZ.AU (Richard A. O'Keefe)
Newsgroups: comp.lang.prolog
Subject: Re: Arity Prolog
Message-ID: <3156@goanna.cs.rmit.OZ.AU>
Date: 5 May 90 03:39:11 GMT
References: <10830@sun.udel.edu> <626@seila.UUCP>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 27

In article <626@seila.UUCP>, johnt@seila.UUCP (john grant) writes
[that Arity/Prolog limits you to 64kbytes for each stack; he says
 that this "seems strange"].

In fairness to Arity, that's a pretty common limitation.

LPA Prolog Professional, for example, has a command line switch /Enn
which controls stack size.  The manual says "It is not possible to
increase evaluation space beyond 64kb, owing to limitations of the
8086 microprocessor family".

ALS Prolog Personal version 1.0 says "At initialisation time, the
system initialises the heap, local stack, and trail each to the same
size.  ... Larger systems allocate each area 64k."  The Professional
version of ALS Prolog also has automatic code swapping, so code size
is effectively unlimited.

The OS/2 operating system is based HEAVILY around the 80286 segmented
architecture.  The 80286 and OS/2 actually make the segment juggling
that is needed to extend the stacks beyond 64k _more_ costly than
the 8086 and MS-DOS do, because of the checking of segment descriptors
and such.  OS/2 is very definitely a 16-bit software architecture.

I can say that Quintus have looked very closely at OS/2 and continue
to keep an eye on it, but a 386 box with UNIX V.3 and VP/ix or Dos/Merge
is in the same general price range as an OS/2 system, and several vendors
(including Quintus) already have Prologs for the 386.